Analysis and Synthesis of Adaptive Neural Elements and Assemblies
Abstract
Between October 1, 1991 and September 30, 1992, progress was made in four areas. First, the capabilities of SNNAP, a general purpose Simulator for Neural Networks and Action Potentials, were enhanced by incorporating mathematical descriptions of intracellular levels of Ca2+ and second messenger systems, which in turn modulate membrane conductances. Second, cellular mechanisms underlying operant conditioning were investigated in simulations of neural networks with biologically realistic properties. In one neural network, a learning rule (activity-dependent neuromodulation), which has been proposed as a cellular mechanism for classical conditioning , was demonstrated to support many features of operant conditioning. A second neural network was developed that simulates the biophysical properties of the neurons and synaptic interactions in a central pattern generator (CPG) underlying aspects of feeding behavior - a behavior that can be modified by operant conditioning. Third, experiments characterized the modulatory actions of transmitters on the synaptic connections and the intrinsic biophysical properties of neurons in the feeding CPG. Fourth, extensions were made to the single-cell model of associative learning by incorporating quantitative descriptions of the modulation of membrane currents by 5-HT.... Learning, Memory, Information storage, Artificial intelligence, Neuronal and neural network computations.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 14, 1992
- Accession Number
- ADA259954
Entities
People
- John H. Byrne
Organizations
- McGovern Medical School